Sequential detection based classifications of RFID tags in three-dimensional space
Systems and methods for sequential detection-based classifications of radio-frequency identification (RFID) tags in three-dimensional space are provided. The methods include modeling a response from RFID tags as a probabilistic macro-channel and interrogating an RFID tag by transmitting a series of packets. Each packet is a transmit symbol and a first series of packet is a transmitted codeword. The method includes receiving, from the RFID tag, a second series of packets that is a received codeword in response to the transmitted codeword and finding a jointly typical transmit and receive codeword across all classes of macro-channels. The method also includes declaring a class of the RFID tag based on a largest likelihood between the transmitted codeword and the received codeword.
1. A method for sequential detection-based classifications of radio-frequency identification (RFID) tags, comprising:
identifying RFID tags at an RFID checkout station of a store by employing an RFID reader;
defining one or more areas adjacent the RFID checkout station as tunnels, each tunnel being a three-dimensional volume including a boundary, a shaping of the boundary of each tunnel facilitated by RF absorbers and RF reflectors;
modeling a response from the RFID tags as a probabilistic macro-channel, the probabilistic macro-channel being an abstraction of a received reply under a given interrogation scenario and specified parameters, the abstraction analyzed as a discrete memoryless channel with a transition probability;
interrogating, by at least one processor device, at least one RFID tag by transmitting a series of packets where each packet is a transmit symbol and a first series of packet is a transmitted codeword;
receiving, from the at least one RFID tag, a second series of packets that is a received codeword in response to the transmitted codeword;
finding, by the at least one processor device, a jointly transmit and receive codeword across all classes of macro-channels; and
declaring a class of the at least one RFID tag based on a largest likelihood between the transmitted codeword and the received codeword, the class being either a first class of an RFID tag within a tunnel or a second class of an RFID tag outside a tunnel formed adjacent the RFID checkout station.
2. The method of claim 1 , further comprising:
determining a likelihood between two codewords based on a distance.
3. The method of claim 2 , wherein the distance is Kullback Leibler (KL) distance defined as:
D
KL
(
f
i
n
,
f
j
n
)
=
-
Σ
y
n
f
i
(
y
k
n
❘
x
n
)
log
f
j
(
y
n
❘
x
n
)
f
i
(
y
n
❘
x
n
)
,
where a channel output y is a vector of received signal attributes, a channel input x a vector that consists of attributes used by the at least one processor device to interrogate the at least one RFID tag, n is a number of observations and i and j are two channels in at least one channel set.
4. The method of claim 1 , wherein the RFID tags are identified within the one or more tunnels in a particular time interval.
5. The method of claim 1 , wherein the RFID tags are identified in different sessions, a session defined as a four-dimensional section marked by a limited time interval in a particular tunnel.
6. The method of claim 1 , wherein a metric for a tunnel and an RFID tag is given as:
M ( T ,∈)=Σ i=1 2 W μ( T,y i 1 ,x i 1 ,y i 2 ,x i 2 . . . ,y i n ,x i n )
where 2 W is a power set of a set W that consists of all n-tuples of elements of a set S where the elements of the n-tuples are always in ascending order, and M is a sum of all metrics for a given electronic product code (EPC) of an arbitrary RFID tag c for tunnel T.
7. The method of claim 1 , wherein the probabilistic macro channel is a function of a phase difference and Doppler shift between a transmit signal and a receive signal.
8. The method of claim 1 , further comprising:
computing a likelihood between a transmitted signal and a received signal for each class sequentially as symbols are received.
9. The method of claim 8 , wherein computing the likelihood depends on an entire series of symbols that have been received.
10. The method of claim 8 , wherein computing the likelihood further comprises:
using a time dependent function.
11. The method of claim 8 , wherein the likelihood for each class is compared to a threshold to declare if the at least one RFID tag belongs to the class.
12. The method of claim 8 , wherein a combination of two likelihoods and μ 2 defined as:
μ 1 ⊕μ 2 =μ 1 +μ 2 −μ 1 *μ 2 .
13. A computer system for sequential detection-based classifications of radio-frequency identification (RFID) tags, comprising:
a processor device operatively coupled to a memory device, the processor device being configured to:
identify RFID tags at an RFID checkout station of a store by employing an RFID reader;
define one or more areas adjacent the RFID checkout station as tunnels, each tunnel being a three-dimensional volume including a boundary, a shaping of the boundary of each tunnel facilitated by RF absorbers and RF reflectors;
model a response from the RFID tags as a probabilistic macro-channel, the probabilistic macro-channel being an abstraction of a received reply under a given interrogation scenario and specified parameters, the abstraction analyzed as a discrete memoryless channel with a transition probability;
interrogate at least one RFID tag by transmitting a series of packets where each packet is a transmit symbol and a first series of packet is a transmitted codeword;
receive, from the at least one RFID tag, a second series of packets that is a received codeword in response to the transmitted codeword;
find a jointly transmit and receive codeword across all classes of macro-channels; and
declare a class of the at least one RFID tag based on a largest likelihood between the transmitted codeword and the received codeword, the class being either a first class of an RFID tag within a tunnel or a second class of an RFID tag outside a tunnel formed adjacent the RFID checkout station.
14. The system as recited in claim 13 , wherein the processor device is further configured to:
determine a distance between two codewords based on a Kullback-Leibler (KL) distance.
15. The system as recited in claim 13 , wherein the RFID tags are identified within the one or more tunnels in a particular time interval.
16. The system as recited in claim 13 , wherein the RFID tags are identified in different sessions, a session defined as a four-dimensional section marked by a limited time interval in a particular tunnel.
17. The system as recited in claim 13 , wherein a metric for a tunnel and an RFID tag is given as:
M ( T ,∈)=Σ i=1 2 W μ( T,y i 1 ,x i 1 ,y i 2 ,x i 2 , . . . ,y i n ,x i n )
where 2 W is a power set of a set W that consists of all n-tuples of elements of a set S where the elements of the n-tuples are always in ascending order, and M is a sum of all metrics for a given electronic product code (EPC) of an arbitrary RFID tag ϵ for tunnel T.
18. The system as recited in claim 13 , wherein the probabilistic macro channel is a function of a phase difference and Doppler shift between a transmit signal and a receive signal.
19. The system as recited in claim 13 , wherein the processor device is further configured to:
compute a likelihood between a transmitted signal and a received signal for each class sequentially as symbols are received.
20. A computer program product for sequential detection-based classifications of radio-frequency identification (RFID) tags, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to perform the method comprising:
identifying RFID tags at an RFID checkout station of a store by employing an RFID reader;
defining one or more areas adjacent the RFID checkout station as tunnels, each tumid being a three-dimensional volume including a boundary, a shaping of the boundary of each tunnel facilitated by RF absorbers and RF reflectors;
modeling a response from the RFID tags as a probabilistic macro-channel, the probabilistic macro-channel being an abstraction of a received reply under a given interrogation scenario and specified parameters, the abstraction analyzed as a discrete memoryless channel with a transition probability;
interrogating, by at least one processor device, at least one RFID tag by transmitting a series of packets where each packet is a transmit symbol and a first series of packet is a transmitted codeword;
receiving, from the at least one RFID tag, a second series of packets that is a received codeword in response to the transmitted codeword;
finding, by the at least one processor de vice, a jointly transmit and receive codeword across all classes of macro-channels; and
declaring a class of the at least one RFID tag based on a largest likelihood between the transmitted codeword and the received codeword, the class being either a first class of an RFID tag within a tunnel or a second class of an RFID tag outside a tunnel formed adjacent the RFID checkout station.